MD+DI Online is part of the Informa Markets Division of Informa PLC

This site is operated by a business or businesses owned by Informa PLC and all copyright resides with them. Informa PLC's registered office is 5 Howick Place, London SW1P 1WG. Registered in England and Wales. Number 8860726.

Sitemap


Articles from 1996 In January


Clinical Trials: An Introduction

Kshitij Mohan and Harold E. Sargent

Throughout most of human history, the advance of medical practice has depended largely on accidental discoveries and observational data. Often, treatments or therapies with unmistakably positive effects on patients have been discovered by chance and subsequently adopted into medical practice. In other cases, conjectures or theories have been put forward by the leading authorities of the day, and have found their way into medical practice without any sort of controlled experimentation. For instance, the practice of bloodletting persisted well into the 18th century before new understandings of human physiology drove it into disuse.

Until our own century, instances in which controlled experiments led to the discovery of medical benefits have been extremely rare. The experiments that led to the discovery that citrus juice could be used to prevent scurvy in sailors offer a notable exception to the rule. While other fields of human endeavor were benefiting from the scientific method that developed during the Enlightenment of the 18th century, application of that method to the field of medicine lagged far behind.

Nowhere is this more true than in the area of clinical trials---controlled experiments designed to evaluate the safety and efficacy of one or more medical treatments using human beings as the patients or subjects. While the modern clinical trial embodies the principles of the scientific method, development of techniques for conducting clinical trials came long after that method had begun to be applied to medicine in general. In fact, only since World War II has the controlled, randomized clinical trial come into widespread use in the medical field.

The first controlled clinical trials in the United States focused on providing evidence that the products being tested were safe for human use. It was not until 1970 that clinical trials were required to address both safety and efficacy. Today, the range of questions being addressed in clinical trials includes not only product-related issues, but also quality-of-life measurements and, most recently, economic analyses.1

The codification of clinical trial methodologies into laws and regulations has had a dual effect on the advancement of such methodologies. On the one hand, it has made more prevalent the use of clinical trials to evaluate medical products prior to their introduction into the marketplace; on the other, it has sometimes resulted in a ritualistic and inappropriate reliance on clinical trials for the answers to questions that they are not capable of addressing. These effects have become especially apparent in relation to the clinical trials used to examine medical devices or complex therapies that combine devices, drugs, and new medical procedures. Over the past several years, a variety of trends--including the increasing complexity of medical devices, ever-growing requirements for conducting clinical trials, and more-powerful demands for information related to cost-effectiveness and clinical efficacy--have combined to challenge the limits of the clinical trial as an objective method of scientific inquiry.

THE PROTECTION OF HUMAN SUBJECTS

It is ironic that the irrational use of clinical trial methodologies--a rational scientific tool intended for the human good--has made it necessary to establish social restraints in order to protect the human subjects who participate in such trials. Unfortunately, however, the 20th century has witnessed all too many instances in which the deliberate misuse or ignorant use of clinical trial methods has made such restraints necessary. The bestiality displayed by the clinical researchers of the Third Reich in the concentration camps of World War II, and the callousness of those who studied syphilis in unknowing and untreated African American men in Tuskegee, AL, are only a couple of the better-known examples. Other examples include the alleged forcible use of prisoners or other institutionalized persons as trial subjects, particularly in totalitarian societies, and the careless or callous disregard for human subjects in radiation experiments conducted by the United States over a number of decades.

In reaction to such occurrences, several attempts have been made to define justifiable boundaries between the rights of the individual and the benefits that might accrue to society through the scientific information developed in clinical trials. An early example of such an effort is the Nuremberg Code, which was a set of criteria established in 1947 as a means of adjudicating the Nazi researchers who experimented with human beings in concentration camps during World War II. But the Nuremberg Code is by no means the earliest of such works.

The Hippocratic Oath. The preeminent code of ethics for physicians, the Hippocratic Oath has existed since the fourth century BC. At its 1948 meeting in Geneva, the World Medical Association drafted a modern version of the oath, which is often referred to as the Declaration of Geneva. This version was further amended in 1968.

In its modern form, the Hippocratic Oath has two main parts. The first deals with the duties of the physician to his or her teachers and his or her role in transmitting medical knowledge to others. The second part is a summary of general principles of medical ethics as they relate to the treatment of patients. The scope of the oath is thus limited to clinical practice, and it does not directly touch on ethics as they might relate to research. Since clinical research cannot always be separated from the treatment of patients, however, the ethical guidelines present in the oath may also be considered to apply to the physician's treatment of research subjects. Hence, the oath's first dictum to "do no harm" is also a key principle of ethics as they relate to clinical research.

Declaration of Helsinki. In 1964, the World Medical Association drew up a code of ethics that was intended to address the unique circumstances surrounding human experimentation. Known as the Declaration of Helsinki, this document was subsequently revised by the 29th World Medical Assembly in Tokyo in 1975, and by the 35th World Medical Assembly in Venice in 1983.

In its current form, the document consists of three parts. The first outlines 12 basic principles that should guide all clinical research. These include the principles that clinical research should be conducted in conformity with currently accepted scientific principles; that the research objectives should be balanced against the inherent risk to the subject; and that research subjects are entitled to informed consent.

The second part lists six principles for medical research when combined with patient care. These include protection of the patient's right to participate or not participate in a study without impact on the physician-patient relationship; reiteration that the interests of the patient must come first; and the statement that combining medical research with patient care is justified only to the extent that the research offers potential diagnostic or therapeutic value to the patient.

The third part of the declaration outlines four principles for medical research being conducted with healthy human volunteers. These principles reiterate that the interests of the subject always take precedence over the interests of science.

The Belmont Report. The ethical paradigm that most closely defines current practices is contained in an elegant exposition known as the Belmont Report, which was published in 1978 by the National Commission for the Protection of Human Subjects of Biomedical and Behavioral Research.2 This commission took as its scope the protection of human subjects in research rather than in the normal practice of medicine or even in the experimental practice of medicine. The purpose of practice is solely the benefit of a patient, that of research is to test a hypothesis to enhance the general public good. To ensure that all subjects involved in such research activities would be protected, the commission intentionally defined its scope to include any activity that included a research component. In the course of its work, this group commissioned a number of scholarly papers from which emerged three basic ethical principles that underlie many of today's seemingly bureaucratic requirements: respect for persons, beneficence, and justice.

Respect for persons relates to the right of a person to be the ultimate decision maker regarding his or her participation as a subject. To enable the individual to exercise such a right, it is necessary that the individual be informed about all the implications of his or her decision and that it be arrived at freely. The right to an informed and free decision can be compromised if the individual has diminished autonomy, that is, if he or she is not capable of being fully informed, as in the case of mental patients, children, or others with diminished capability to comprehend sufficiently the implications of their decisions. Diminished autonomy also applies to persons under duress or under other circumstances that restrict their ability to arrive at a decision freely. Such is the case with prisoners or subjects of totalitarian states. While complicated situations may sometimes require a judgment call, such judgments should be guided by the principle that the decision to participate in a clinical trial should be both free and informed, and special protection should be exercised on behalf of individuals for whom either of those capacities is diminished.

This principle of respect for persons has led to today's requirements and practices related to informed consent. The study subject must receive sufficient information, and in a manner that he or she can comprehend. Thus, not only does information about the study need to be in the subject's language, but it must be at a level that allows comprehension. Information on all risks and benefits must be adequate to enable a reasonable person to make a reasoned decision. Consent needs to be totally voluntary--without social, economic, or other forms of coercion. For those with inadequate ability to fully comprehend--such as children, the mentally handicapped, the terminally ill, or the comatose--a surrogate, such as a parent, a spouse, or someone authorized or recognized as acting in the subject's best interests must be the recipient of the information and the grantor of the consent.

The Belmont Report defines the principle of beneficence as not doing harm, and maximizing possible benefits while minimizing possible harm. The report borrows the Hippocratic dictum "do no harm," and interprets this to mean that harm to even a single person cannot be justified--no matter how much good it does for others. Thus, in a clinical trial, there should be the promise of benefit to the subject as well. Preclinical studies, including animal studies where necessary, also help in minimizing any possible risk to the subject. At the same time, the trial should be designed in such a way that its overall benefits to society are maximized.

The principle of beneficence has led to today's requirements for assessing the risks and benefits involved in a study, minimizing the former and maximizing the latter. Thus, preclinical and animal studies are required, and the general tenet applies that if any scientific information can be gathered by a method other than using human subjects, that method should be applied before conducting clinical trials. Brutal or inhumane treatment of subjects or unnecessary risk are not allowed, and appropriate documentation of the risks and benefits, their appropriate peer review by an objective group such as an institutional review board (IRB), and the inclusion of the information in informed consent forms are required.

Another requirement derived from the concept of beneficence is that research should be conducted in a scientifically valid manner, using a protocol that meets the norms of good science, so that the probability of the research being successful is high and therefore the benefit to society is maximized. Sloppily conceived or conducted research does not justify the exposure of subjects to any risks. Also, with an appropriately conducted clinical trial, information regarding risk or benefit can emerge that could make it unethical to continue as originally planned. If there is clear evidence of inadequate safety or effectiveness, the trial would have to be aborted. On the other hand, if there is evidence of high effectiveness, it may be unethical to restrict the availability of the therapy only to research subjects.

The final principle discussed in the Belmont Report is that of justice. Justice demands that research be conducted fairly, and that both its burdens and benefits be shared evenly. Thus, exposure to the risks involved in clinical research should not be borne by any particular group, such as prisoners, or by groups that are economically or socially disadvantaged. If the benefits of the research flow to society in general, a fair cross section of society should also bear the risks.

In current practice, the concept of justice has led to the creation of specialized requirements related to the selection of research subjects. Generally speaking, special care and safeguards must be used for any study that seeks to include prisoners, minorities, or the institutionalized, or for any study that includes children early on. On the other hand, justice may also require that some groups, such as women, not be excluded. This is to ensure that clinical research and the subsequent advancement of therapies is applicable to the greatest possible number of people.

Other Documents. The use of human subjects in clinical research is the subject of many other government documents, including a set of guidelines established by the U.S. Department of Health, Education, and Welfare in 1971, and made into regulations in 1974. Naturally, FDA's regulations and guidelines related to the clinical investigation of new drugs and devices are numerous. Dealing with their requirements has launched a whole profession of bioethicists and an entire industry of contract research organizations (CROs) whose purpose is to design and implement clinical trials in a socially responsible manner. Recent findings on bioethical lapses in the human radiation research funded by the U.S. government since World War II indicate that the last word on the protection of human subjects in research will continue to be rewritten.3

To be continued with: THE SCIENCE OF CLINICAL TRIALS

The Gamma Radiation Tolerance of Polypropylene: Measurement and Enhancement

Medical Plastics and Biomaterials Magazine | MPB Article Index

Originally published January 1996

 

ROBERT C. PORTNOY

Although many tests are currently being used to evaluate the high-energy radiation tolerance of polypropylene, most have little in common with the polymer's performance in its intended application. Conversely, setting up application-specific tests can be expensive and time-consuming. This paper describes a test that can be conveniently performed using ASTM parts, but which has relevance for medical device and labware applications. The method involves an extension of the ASTM flexural-modulus determination and provides a quick and reliable method for rating radiation tolerance without the use of complicated parts or procedures. The use of this test for the validation of several approaches to enhancing polypropylene radiation tolerance is also discussed.

INTRODUCTION

Polypropylene is entering a golden period of applications in medical devices and device and drug packaging. The wide variety of homopolymers, random copolymers, and impact copolymers commercially available at very economical prices is rapidly stimulating the use of polypropylene in medical fields. Especially important to the new popularity of these plastics are their clarity and their ability to withstand all major methods of sterilization. Free of haze and the sensitivity to high-energy radiation that plagued polypropylene in the past, the new formulations are in high demand for uses previously reserved for glass or other, more costly, plastics. Considering the range of capabilities represented by the many polypropylenes now offered to the medical industry, it is absolutely necessary for the medical product designer to understand the factors affecting the radiation tolerance of the resins as well as the methods used to determine tolerance levels. Such knowledge is especially critical given the trend toward high-energy sterilization of medical devices and away from sterilization by ethylene oxide.1

Conventionally stabilized polypropylenes are not suitable for sterilization by high-energy radiation because of the severe embrittlement and discoloration that occur immediately in the plastic after sterilization and worsen with aging. Whereas the embrittlement of the plastic after irradiation is an inherent property of the polymer, the discoloration is caused by reaction products of the phenolic antioxidants normally included in standard polypropylenes.1

There are, however, several techniques in the design of propylene polymers and their formulations that remedy these problems and yield resins suitable for irradiation at dosages up to 50 kGy. Early radiation-tolerant polypropylenes were homopolymers stabilized with small quantities of phenolic antioxidants and large quantities of sulfide-diester secondary antioxidants. These materials did discolor slightly after irradiation.

The modern resins that are most successful in withstanding irradiation exhibit reduced crystallinity and narrow molecular-weight distribution, are formulated with hindered amine light stabilizers (HALS), and contain no discoloring phenolic antioxidants. Ethylene-containing random copolymers are useful substrates for building radiation-tolerant formulations, as are homopolymers with low isotacticity and homopolymers to which hydrocarbon oils or greases have been added. The HALS are, by themselves, noncoloring in polypropylene, but they interact with phenolic antioxidants to produce extremely deep yellow colors after irradiation. Therefore, when HALS are used in a polypropylene formulation, the phenolic antioxidants must be scrupulously avoided.1-6

Over the years, numerous methods have been developed to investigate and evaluate the effects of ionizing radiation--such as that from gamma and electron-beam sources--on polypropylene. The types of tests can be categorized broadly as spectroscopic, macroscopic, and end-use-specific. The spectroscopic type, which includes analytical and compositional analyses, involves such methods as electron spin resonance spectroscopy, infrared spectroscopy, and chemiluminescence.7­9 Although these methods have given great insight into the physics and chemistry occurring during and after irradiation, the experimentation can be costly, extensive, and complicated. Also, because of the complex nature of polymeric systems, the results can be equivocal. Even though correlations of spectroscopic results with specific physical properties have been made in certain cases, they are seldom used in place of direct physical property measurements.10,11

The macroscopic investigations comprise all gross physical and rheological property measurements--for example, flexural modulus, notched Izod and Gardner impact, tensile elongation, and melt-flow rate.12 Although these measurements can show the effects of radiation, some have very little, if any, correlation with the end-use application. There is also quite a lack of agreement among polymer suppliers and users as to which tests are most relevant and the specific conditions under which they should be conducted.

Although the end-use-specific methods--tests such as syringe-flange bending, luer-tip breaking, syringe-barrel crushing, and tube centrifugation13--are generally the best indicators of actual field performance, they also have some drawbacks, especially for the resin producer. One of the biggest disadvantages is the expense of acquiring and operating specialized equipment for both parts fabrication and testing. The resin supplier is obliged to set up a separate test method for each end use and according to each customer's protocol, necessitating extensive testing with a broad spectrum of equipment.

The flex-to-failure method described here and in a previous publication14 is an offshoot of the general bending concepts related by Apostolou,15 the specific syringe-flange bending method of Williams, et al.,16 and the notched, three-point flexural method of Narisawa and Ishikawa17 for investigating crazing in semicrystalline thermoplastics. The test proposed here is an approach to screening resin tolerance to high-energy radiation--especially gamma and electron-beam radiation--as it specifically relates to performance concepts and characteristics sought in the medical field. The method, which involves an extension of flexural stress-strain testing, allows for the convenience and universality of ASTM testing without the use of complicated parts or procedures. In this test, sections from an ASTM tensile test bar ("dog bone") used for the determination of flexural modulus are flexed to failure, either brittle or ductile, whichever occurs first. The test's peak load, deflection, and energy-consumed-to-failure values have been used by Exxon Chemical Co. in the past as gauges of radiation tolerance. As researchers gained experience in this testing, it was determined that the deflection at peak flexural load was the test's most valuable measure of polypropylene embrittlement after irradiation. The method is sensitive to radiation-induced polymer changes and provides a quick, reliable, and relevant screening test, representative of actual medical device failure modes.

This paper presents the results of a number of related experiments demonstrating the power of the flex-to-failure test as part of a total screening regimen in the evaluation of some modern methods of stabilizing polypropylenes to high-energy radiation.

EXPERIMENTAL

Four different new experiments are documented in this paper. In each of these, the flex-to-failure technique was used to assess the response of polypropylene formulations to high-energy radiation. Other common test methods were also employed to provide a framework for understanding the effect of the radiation on the various formulations and to provide some perspective on the value of the flex-to-failure technique.

In all four experiments, polypropylene formulations were compounded and pelletized, then molded into ASTM test parts in a family mold. Only Gardner disks (88.9 mm diam x 3.18 mm thickness) and tensile bars (165 mm length x 12.7 width x 3.18 mm thickness) were required for the tests performed in this work. Parts from each resin were irradiated by a Co60 gamma source at approximately 10 kGy/hr to each of the indicated total dosages (see Figures 1­12). The parts were then aged at 60°C for 21 days and the testing carried out as soon as possible after the aging period. The tests performed were the determination of Gardner impact strength (ASTM D 3029-84), tensile elongation at break (ASTM D 638-87b), and flex-to-failure (ASTM D 790-86 extended past peak load), all at 23°C. The latter procedure simply extended to higher strain the determination of flexural modulus, and comprised the entire flex-to-failure test. The deflection at maximum stress corresponding to either brittle or ductile failure was observed and recorded as the deflection at peak flexural load.

Experiment 1: Effect of Ethylene Content on the Radiation Response of Random Copolymers. Samples 1­4 were identically formulated, 25 dg/min melt-flow rate (MFR) polypropylenes, polymerized with 0, 1.5, 2.2, and 2.8% ethylene comonomer, respectively. The formulations contained both hindered amine light stabilizer and nucleating clarifier as well as other standard ingredients used in Exxon's commercial medical device formulations.

Experiment 2: Effect of Mobilizing Oil on the Radiation Response of Propylene Homopolymer. Samples 5 and 6 were identical 25 dg/min MFR propylene homopolymers containing the standard additive package, except that an additional 3% of a paraffin oil was compounded into Sample 6.

Experiment 3: Comparison of the Protective Effect on Irradiated Polypropylenes of Two Different Hindered Amine Light- Stabilizer Packages. Samples 7 and 8 were two 1.5% ethylene random copolymers with 25 dg/min MFR. Sample 7 was stabilized with a high level of Tinuvin 622 (Ciba Geigy Corp., Hawthorne, NY) while Sample 8 was stabilized with a similar amount of Tinuvin 770. Both resins were nucleated and clarified.

Experiment 4: Effect of Reformulation on a Successful Radiation- Tolerant Random Copolymer. Sample 9 was Exxon's commercial, medical device resin PP 9074MED. Sample 10 was the same polymer, stabilized in exactly the same way, but nucleated with Millad 3988 instead of Millad 3940 (Milliken Chemical Div., Spartanburg, SC) and containing other small additive modifications.

RESULTS AND DISCUSSION

The results of the four experiments are presented in graphical form in Figures 1­12.

The Flex-to-Failure Test Profile. In the flex-to-failure test, four different stress-strain profiles have been observed. Figure 13 shows the characteristic ductile failure mode. As the sample is flexed in the three-point bending configuration, the stress increases with deflection until ductile failure occurs, corresponding to the peak load of the profile. As expected, most nonirradiated polypropylenes will exhibit this type of profile. Lower peak loads are generally associated with more flexible samples.

From past studies conducted by the author, it is known that four different failure modes can be observed in the flex-to-failure test. In order of increasing brittleness, the sample may exhibit ductile failure only (Figure 13), ductile failure followed by brittle failure (Figure 14), stepwise brittle fracture without ductile failure (Figure 15), or simple brittle fracture without ductile failure (Figure 16). The stepwise brittle failure is rare, since it occurs only at a very specific balance of ductility and brittleness and may represent a sample that is not uniform throughout its cross section at the point of application of the stress. The more ductile the polypropylene sample, the greater is the deflection until failure, up to a limit of about 11.0 mm. The numerical results recorded here correspond to all of the qualitative failure modes except the stepwise brittle fracture.

The Effects of Formulation Differences on the Radiation Resistance of Polypropylenes. While the experiments described here in no way constitute an exhaustive study of the factors affecting the behavior of polypropylene exposed to high-energy radiation, they do provide some illustration of the primary influences on the radiation resistance of this thermoplastic.

Figures 1 through 3 clearly show the increase in ductility imparted to polypropylene by random ethylene comonomer, an effect that increases monotonically with comonomer content. This increase in ductility is evident in nonirradiated material as well as in the irradiated samples through the improvements in the tensile elongation and impact strength of the materials. The evidence of these changes in nonirradiated samples is valuable to an understanding of the effect of ethylene comonomer on the basic physical properties of resins. However, it complicates comparisons of embrittlement of irradiated materials by giving different starting points for each material. The deflection at peak flexural load is remarkably similar for all four of the ethylene levels studied in Experiment 1. While much less information about the ductility of nonirradiated resin is available from these results, the embrittlement that occurs upon irradiation is much easier to compare for the various examples.

Despite their differences, all of the tests point to the same conclusion: that ethylene comonomer is a powerful stabilizer of polypropylene to the effects of high-energy radiation. Random copolymers with levels of ethylene greater than about 2% and further stabilized with a suitable additive package should be useful in applications requiring doses even higher than 30 kGy.

The results for samples 5 and 6 showed a similar enhancement of ductility produced by the addition of 3% of a paraffin oil to a HALS-stabilized propylene homopolymer. All three of the test methods used were complementary in giving an overall picture of the effect of the oil additive. At 3% of the formulation, the oil extended the ductility of the irradiated resin in a manner very much like that of ethylene comonomer at 3% (Sample 4, Figures 1­3), but had much less effect on the starting properties of the nonirradiated material.

Testing on samples 7 and 8 verified the marginal superiority of Tinuvin 770 as a protectant for irradiated polypropylene compared with a similar concentration of Tinuvin 622. These results from Experiment 3 also showed the value of using all three of the test methods to obtain a comprehensive picture of the behavior of irradiated polypropylene. The Gardner impact strength of the irradiated resin appeared to be most susceptible to damage by irradiation and likewise most sensitive to differences between the two stabilizers, even at dosages below 25 kGy. Tensile elongation at break was less affected, and showed a difference between the stabilizers only above 25 kGy. The deflection at peak flexural load was least affected by the irradiation, and was unchanged for both resins up to 50 kGy. The slight inferiority of the sample stabilized with Tinuvin 622 became evident in the flexural test only at 75 kGy, when the deflection first decreased significantly.

Samples 9 and 10 further emphasized the differences in character among the three tests. For these two nearly identical samples--which differed only in auxiliary additives and not in stabilizers or nucleating agent--very small differences in embrittlement behavior after irradiation were expected. This was in fact observed, especially in the results of the tensile and flexural testing. The flexural testing, particularly, indicated the similarity of the resins, even with 75-kGy irradiation. As always, the Gardner impact strength seemed to be more sensitive than the other tests to very minor differences between materials.

Comparison of the Three Testing Methods. Although the three test methods were all useful in investigating the embrittlement of irradiated polypropylene samples, it is valuable to review their particular strengths and weaknesses as illuminated by this study.

Gardner impact strength was extremely sensitive to very minor differences in resin formulation, with or without irradiation. The minimum level of impact strength representing ductility is, in general, difficult to establish because of variation in onset of brittleness for different polypropylene types. When the data obtained in this study is compared with the established utility of commercial formulations such as Sample 4, the Gardner results appear to provide an overly pessimistic prediction of resistance to embrittlement. Strength losses begin to occur with small doses of radiation, but the correspondence between these quantitative losses and the onset of undesirable brittleness is not easy to detect because this relationship is different for different materials.

The Gardner technique also provides a crude qualitative indication of the embrittlement of each sample. Ductility even in a failed sample is indicated by an otherwise intact specimen with only a smooth hole punched through it by the falling tup. At the other extreme, a brittle sample, upon failure, is shattered by the falling tup. Observations of this type were not recorded in this work, but these kinds of additional insights into the behavior of irradiated resins can be very valuable, especially in light of the difficulty of judging embrittlement based solely on strength results.

The tensile elongation at break gives one a more balanced view of the resistance to embrittlement of irradiated polypropylenes: the numerical response to irradiation seems to more accurately track the general utility of the materials in actual applications. But, as in the Gardner impact test, the effect of sample molecular characteristics and impact modifications on elongation makes it difficult to compare a variety of materials or to establish a single value of elongation that represents the demarcation between ductility and brittleness. And, unlike for a Gardner specimen, it is not so easy to judge the ductility of an elongated sample by the appearance of the test specimen after it is broken.

Deflection at peak flexural load from the flex-to-failure test gives a third perspective on determining ductility. While this measurement seems, at the outset, much less sensitive to irradiation and product stability, it in fact provides a simpler approach to accurately judging resin embrittlement. Basically, nearly all polypropylenes show a deflection of from 9.5 to 10.5 mm when not irradiated. No change in this value is seen at very minor levels of resin degradation after irradiation. Thus, the first decline of this value--even by as little as 0.5 mm--denotes significant degradation and therefore can serve as a useful limiting level of irradiation, storage time after irradiation, or both. The brittle forms of breakage shown in Figures 14­16 predominate in samples with lower deflection-at-peak-flexural-load results.

The differences among these three test methods should not be cause for concern. These particular results cannot be unequivocally extrapolated to actual medical devices, in which design, manufacturing, and use factors play so large a role in determining the upper limits of radiation tolerance. The value of these tests is in the consistency of the relative results. The higher-ethylene copolymers, oil-treated homopolymer, and Tinuvin- 770-stabilized homopolymer were shown to be superior in the screening experiments by each of the three test methods. It is this agreement on the relative order of resistance to embrittlement of samples that makes the use of all three of these complementary methods together such a powerful tool.

CONCLUSION

This work demonstrated several important aspects of the formulating and testing of radiation-tolerant polypropylenes. Numerous factors--including comonomer level, choice of stabilizer, and use of mobilizing oil--can affect the embrittlement behavior of clear polypropylene after irradiation. Measuring the effects of these factors can be done by a variety of simple physical tests. Although the results of the Gardner impact, tensile elongation, and flex-to-failure tests are all complementary, the latter provides the most obvious quantitative and qualitative indicator of resin embrittlement. Differences in the mechanisms of applied stress and sensitivity of these tests make it valuable to use more than one type to obtain a complete understanding of the response of a material to irradiation. If possible, the testing regimen should include a method that mimics the stress experienced in actual use by the device that is to be fabricated from the material. Quite often, this is a bending stress, making the flex-to-failure method an important component of any complete regimen for screening the radiation response of polypropylene.

ACKNOWLEDGMENTS

The author thanks Claude Watkins for his work in performing the experiments discussed in this paper. Without his careful attention to detail, the study could not have been completed successfully. The author is also grateful for the contributions of V. R. Cross, coauthor of a previous paperon which this work is based.14

REFERENCES

1. Portnoy RC, "Polypropylene for Medical Device Applications," Med Plas Biomat, 1(1):43­47, 1994.

2. Williams J, Dunn T, and Stannett V, Irradiation sterilization of semi-crystalline polymers, U.S. Pat. 4,110,185, 1978.

3. Williams J, Dunn T, and Stannett V, Semi-crystalline polymers stabilized for irradiation sterilization, U.S. Pat. 4,274,932, 1981.

4. Williams J, Dunn T, and Stannett V, Semi-crystalline polymers stabilized for irradiation sterilization, U.S. Pat. 4,467,065, 1984.

5. Williams JL, and Titus GR, Polyolefin compositions of high clarity and resistance to oxidation, U.S. Pat. 4,845,137, 1989.

6. Toda T, Kurumada T, and Murayama K, "Progress in the Light Stabilization of Polymers," in Polymer Stabilization and Degradation, ACS Symposium Series (280), Klemchuck P (ed), Washington, DC, American Chemical Society, p 37, 1985.

7. Dunn TS, Williams J, Sugg H, et al., "Stability of Gamma-Irradiated Polypropylene. II. Electron Spin Resonance Studies," Adv in Chem Ser, 169:151, 1978.

8. Abdel Kerim FM, Elegrami AM, El-Kalla EH, "Study of the Effect of Gamma Radiation in the IR Spectra of Polypropylene," Isotopenpraxis, 21(1):23­25, 1985.

9. Kadir ZA, Yoshii F, Makuuchi K, et al., "Durability of Radiation-Stabilized Polymers. XIII," Die Angewandte Makromoekulare Chemie, 174:131­140, 1990.

10. Carlsson DJ, Dobbin CJB, Jensen JPT, et al., "Polypropylene Degradation by Gamma Irradiation in Air," in Polymer Stabilization and Degradation, ACS Symposium Series (280), Klemchuck P (ed), Washington, DC, American Chemical Society, p 359, 1985.

11. Mendenhall GD, Agarwal HK, Cooke JM, et al., "Comparison of Chemiluminescence with Impact Strength for Monitoring Degradation of Irradiated Polypropylene," in Polymer Stabilization and Degradation, ACS Symposium Series (280), Klemchuck P (ed), Washington, DC, American Chemical Society, p 373, 1985.

12. Kadir ZA, Yoshii F, Makuuchi K, et al., "Durability of Radiation-Sterilized Polymers. XII. The Effects of Nucleating Agent on the Oxidative Degradation of Polypropylene," Polymer, 30:1425, 1989.

13. Williams J, Dunn TS, Sugg H, et al., "Radiation Stability of Polypropylene," Radiat Phys Chem, 9:445, 1977.

14. Portnoy RC, and Cross VR, "Method for Evaluating the Gamma-Radiation Tolerance of Polypropylene for Medical Device Applications," in Proceedings of the Society of Plastics Engineers (ANTEC XXXVI), Brookfield, CT, Society of Plastics Engineers, p 1826, 1991.

15. Apostolou SF, "Radiation and Plastic: Friend or Foe," presented to the Health Industry Manufacturers Association Conference on Sterilization Methodologies and Materials, Washington, DC, October 1988.

16. Williams J, Dunn TS, Sugg H, et al., "Stability of Gamma-Irradiated Polypropylene. I. Mechanical Properties," Adv in Chem Ser, 169:142, 1978.

17. Narisawa I, and Ishikawa M, "Crazing in Polymers--Volume 2," in Advances in Polymer Science 91/92, Kausch HH (ed), Berlin/Heidelberg, Springer-Verlag, p 354, 1990.

Robert C. Portnoy, PhD, is senior staff scientist at Exxon Chemical Co. (Baytown, TX), where he is involved in the development of products, applications, and markets to further the use of the company's thermoplastics and thermoset elastomers in health care. He specializes in clear and radiation-resistant polypropylenes for injection-molded medical devices.

Continuation of MDDI's January 1996 article, "Clinical Trials: An Introduction"

Back to previous page

Medical devices vary greatly with respect to their type, their population for intended use, and the significance of risk posed by their use. FDA's device regulations recognize these differences by classifying devices into three categories--Class I, II, and III--with Class III representing the highest level of risk of injury to the patient. For the same reasons, FDA does not automatically require clinical trials as part of the approval process for all devices. While the agency is more likely now than it was in the past to request clinical data in conjunction with a 510(k) premarket notification, requirements for clinical trials apply mostly to the technologically innovative devices that reach the market through the premarket approval (PMA) process.

When FDA does require clinical data as part of a product approval submission, it expects that the trials conducted by the manufacturer will make use of the best scientific methods available to answer the specific questions posed for the device. The sections that follow describe some of the issues that commonly arise in the process of designing, conducting, or evaluating clinical research. Subsequent installments of this series will discuss these issues in greater detail and offer specific guidance for manufacturers seeking to navigate these sometimes treacherous waters.

End Points. Every clinical trial should be designed to accomplish a clear study objective in the form of a medical claim for the product. The objective should be as specific as the state of clinical knowledge for that product allows. Hence, for a device undergoing a first feasibility trial, the objective may be related to demonstrating the plausibility of the device concept. For a device more advanced in the development process, the objective may be to compare patient survival times with that device to patient survival times with another currently approved device or treatment regimen. To ensure that the medical claim is entirely accurate, the manufacturer should also identify the intended patient population for the device, including any subpopulations of particular interest. These last considerations lead naturally to the inclusion and exclusion criteria for the study.

Variables. For every clinical trial, the manufacturer should identify all variables that will be monitored as part of the study. Two broad classes of variables are common. Prognostic or baseline variables are collected prior to treatment intervention, and are used as possible covariants in explaining or interpreting a patient's response following intervention. Outcome variables are the primary end points on which assessment of the safety and efficacy of the device will be based. The measurement scale for outcome variables can be continuous (e.g., height, weight, red blood cell count, survival time) or categorical (where the response is classified into one of a finite number of categories). Whenever feasible, continuous-scale responses are preferable because they can provide more information per patient, and thereby reduce the total sample size required to answer the question of interest. Even with continuous-scale responses, appropriate steps should be taken to minimize the potential for measurement bias and to maximize the precision and accuracy of the measurement methods.

Trial Design. The experimental design used for a clinical trial should be that which is most appropriate for the comparisons being made. The two most common types of design for comparing devices or treatment regimens are parallel group studies in which patients are assigned to only one device or treatment regimen, and crossover studies in which each patient receives more than one device or treatment regimen sequentially during his or her time in the clinical trial. Practical considerations such as the time required to complete the trial or the possibility of carryover effects will often determine which design is most appropriate.

The type of control used for the study may also have a determinant effect on its experimental design. Since the control group usually forms the primary basis for a study's comparisons and inferences about the safety and efficacy of the device, selection of the type of control to be used requires careful consideration. The control most commonly used in device clinical trials is an active control that corresponds to a currently accepted device or treatment regimen; placebo controls are relatively rare in device trials. Another alternative is historical controls, which are sometimes used in situations where the overall treatment regimen has not changed markedly over time. From a scientific standpoint, active or placebo controls where patients are randomly assigned to either the study group or the control group provide the most straightforward interpretation of results. In device trials, however, use of placebo controls is often impossible or unethical; even active controls may be difficult to identify. When there is no proven and acceptable treatment available for a life-threatening condition, for instance, the only acceptable control may be historical knowledge of the condition's progress when untreated.

Sample Size. The written protocol of any clinical trial should include a thorough discussion of the sample size planned for the study. An investigation's target sample size should include sufficient overage to accommodate patients who withdraw from the trial or are lost to follow-up. For studies comparing alternative treatment regimens, the manufacturer should derive the sample size required to detect specified differences between the treatments, and should specify the alpha level (risk of concluding that the treatments are different when they are actually equivalent) and power level (probability of detecting a specified difference between the treatments as a statistically significant difference) used for the study.

Statistical power is of particular importance in clinical trials aimed at demonstrating equivalence to an existing treatment regimen. A recent study in the Journal of the American Medical Association found that only 36% of the published randomized clinical trials reported as equivalent had at least an 80% power for detecting a 50% relative difference in response.4 Sample size has a direct effect on the validity of a study: if the sample size is too small to detect even large differences between the test and control groups, one cannot legitimately claim equivalence. The scientific need for large sample sizes, however, needs to be tempered by the realities of available patient populations.

Lies, Damned Lies, and . . .

Medical Device & Diagnostic Industry Magazine | MDDI Article Index

Originally published January 1996

Statistics: I have come to believe that I could spend several hours a day for the next year trying to make sense of FDA review-time numbers and be no wiser for it next January. I don't just mean FDA's own figures--I mean anyone's statistics about agency performance.

This realization first began to overtake me last November, when Congressman Joe Barton's press office sent out a press release decrying what he sees as FDA's misleading use of statistics. What clicked for me wasn't Barton's critique, but Barton's numbers. They didn't appear any more useful, informative, or consistent than FDA's. In fact, they were worse.

It was either Disraeli or Mark Twain (I can never remember which) who said, "There are three types of lies: lies, damned lies, and statistics." Now, statistics have honest uses, or concepts like statistical process control and design of experiments wouldn't work--which they demonstrably do. But statistics will serve any master. Even those persons who cite numbers and percentages with the best intentions can be led astray by their own unwitting biases.

With this situation in mind, I decided that MD&DI should take an objective view of the numbers, if possible. The result is the News & Analysis story, "Who's Right in the FDA Numbers Game?" on page 14 of this issue.

If we came up with any clear answer to that question, it's that no one is right, and no one wins.

Barton's press release charges FDA with cynical manipulation of statistics. Maybe so. But Barton's own use of numbers seems to be on the same level. In his press release, numbers from 1993 are presented as current, no comparison to FDA numbers is made, and no acknowledgment of the improving trend since 1993 is offered.

By comparison, I have to judge FDA's use of numbers to be moderately more complete and consistent. In the November 1 "CEO Letter" from device center director Bruce Burlington, the numbers he presents are clearly meaningful and useful. He offers not just one measure, but three: median review time (the number of days required to decide on half of all applications), average (mean) review time, and the 95th percentile (how long it takes to clear 95% of all applications).

Barton and others have criticized FDA for its use of median rather than mean numbers at various times in the past. But means are easily thrown off by a few outlier applications that take an inordinate amount of time to review. If I want to bash FDA, I'll go with the mean. But if I want a representative view of its performance, give me the median.

For all the clarity and comprehensiveness of Burlington's handling of statistics, he seems to engage in manipulation as well. Although he supplies total review times in one table appended to his letter, his text discusses only FDA review times. In other words, the figures he cites don't really tell how long it takes from receipt of an application to the decision, because they don't include "the time the document was on hold awaiting additional information from the manufacturer."

Come now. Who outside of FDA really cares about FDA time? What really matters is how long it takes to get a product on the market. The figures FDA uses should be based on this simple reality.

So I propose a truce in the statistics war. Let FDA, industry, and Congress get together and settle on one uniform measure for judging FDA performance. The benchmarks for decision times are clear: 90 days for a 510(k) and 180 days for a PMA--total days, not FDA days--as specified by law. All that needs to be determined is the methodology.

Failing this unlikely event, all parties would be wise to restrain their claims about the meaning of the numbers. A statistic, after all, may well be a Class III lie.

John Bethune

Continuation of MDDI's January 1996 article, "Clinical Trials: An Introduction"

Back to front page of story

Over the past few years, there has been much controversy and upheaval in the regulation of medical devices in the United States and, to a lesser extent, in Europe. The result has been a variety of emerging structural and directional changes in the clinical trial arena that bear not only watching, but active participation by the regulatory, clinical, patient, and industry communities.

Drug-Model Science. At FDA, the ethical and organizational controversy that arose over the agency's regulation of generic drugs raised concerns about the rigor of other FDA processes--including the product approval process at its Center for Devices and Radiological Health (CDRH). Shortly afterward, the breast implant controversy seemed to indicate, to some individuals, that the practice of science at CDRH was lacking the physician's perspective. These controversies, together with the appointment of Commissioner David Kessler and the press of congressional investigations, produced a sudden shift in the culture of FDA's device center. Traditionally a bastion of engineers and physical scientists where physicians acted primarily as scientific and medical reviewers, CDRH underwent a rapid change to a physician-managed structure resembling that of FDA's drug center.

In 1992, to investigate some of the perceived concerns regarding the quality of science in the device area, a Committee for Clinical Review was formed and placed under the direction of Robert Temple, director of the Office of Drug Evaluation I at FDA's Center for Drug Evaluation and Research (CDER). Temple had then, and still has, an outstanding record as a scientific conscience of the drug approval process.

The Temple committee reviewed 29 device applications and pointed out a number of deficiencies in their use of scientific evidence.6 Although many of these deficiencies were legitimate, members of the device industry complained that they were valid only when viewed from the perspective of the drug approval process. Industry argued that many of the findings of the Temple Report were not deficiencies at all, but were due to the basic differences between devices and drugs.

To be sure, there are significant differences between drugs and devices. The law and burden of proof for drugs differ from those for devices, and FDA's drug approval process is significantly longer and more elaborate than the approval process for devices. The device law was developed later and included lessons learned from the burdens of the drug process. For example, the medical device regulations define valid scientific evidence much more broadly than the equivalent regulations for drugs do, and thereby allow the device researcher much greater flexibility than is permissible under the drug law and approval process.

The nature of drug research also differs from that of device research. Studies of complex devices may involve equally complex procedures and other concomitant treatments that prevent researchers from using controls to the same extent that they would do in a drug trial. For instance, controls for orthopedic or cardiovascular implants must sometimes be based on historical experience. Equally daunting is the challenge of locating a patient population that can participate in a device trial. For some conditions, the available populations of patients are so small that it would take a very long time to make up a statistically significant sample for a trial.

Without acknowledging these differences, the Temple Report criticized the existing product approval practices at FDA's device center, and effectively established a frame of reference for the new philosophy at FDA. The only problem was that the new philosophy was clearly the drug approval philosophy. In the hands of highly experienced and discriminating scientists, the new direction might have been more manageable. In the hands of a reviewing organization that had been traumatized by seemingly unnecessary changes and brutal congressional second-guessing, it became an excuse to impose new and unnecessary burdens in the name of "good science." Among the many disastrous effects of these political and philosphical changes is the catatonia in the product approval process that has affected the device industry for the past few years.

The controversy over the validity of these different scientific philosophies continues to affect industry. One result has been that manufacturers are moving more of their device-related R&D--including clinical research--to sites outside the United States.7 Sadly, a political phenomenon--the 1994 takeover of Congress by the Republicans--has influenced action on this issue more than all the scientific debate that preceded it. Today, the pendulum at FDA seems to be swinging back to a more moderate approach.

Clinical Utility. Another issue that is influencing clinical research in the United States is that of clinical utility. At its most simplistic level, the issue is whether proof of efficacy relates only to the functional efficacy of the product or requires an outcome-based demonstration of clinical benefit or clinical utility. For example, is it enough to show that a cholesterol-lowering therapy can effectively reduce cholesterol, or does one have to show that such a reduction also provides a proven clinical benefit?

Although FDA has historically allowed some drug and device indications that were clearly based on functional efficacy, in the past few years it has shown a preference for a more rigorous showing of clinical utility. In doing so, the agency has drawn criticism that it is venturing into the regulation of the practice of medicine. This inflammatory argument continues, but answers may yet be found in moderate approaches to the issues. One resolution might be for FDA to consider acceptable surrogate end points that demonstrate functional efficacy, so long as current medical opinion (rather than rigorous proof) considers them to be clinically significant. Such a solution would be particularly applicable to diagnostic products.

Relative Efficacy. If a recent statement published by FDA in the Federal Register is any indication, the question of relative efficacy is another emerging issue that may affect device studies.8 While seemingly innocuous, the statement suggests that FDA intends to broaden its requirements for proof of a device's safety and effectiveness to include a determination of relative efficacy. If this interpretation were to be accepted, labels that now must include adequate instructions for use might also be required to include information comparing the efficacy of the product to that of alternative therapies. In turn, such a requirement would compel manufacturers to conduct multiarm trials in order to compare their new devices with all other competing drug or alternative therapies. Even if that is not the explicit requirement at present, FDA's statement appears to indicate an alarming policy trend.

Leanings in this direction are especially worrisome in light of recent political efforts to influence federal health-care reimbursement policies by involving FDA in the evaluation of outcomes research. So far, the agency has resisted this pressure, but one can envision an integrated system in which outcomes research would become part of the regulatory requirements for product approval. Such a system would require a host of new regulatory metrics for such areas as quality of life, cost comparison, and cost-effectiveness. Efforts to include the fledgling science of outcomes research into a legally mandated criminal code such as the Federal Food, Drug, and Cosmetic Act could well become the next big nightmare of the medical device industry.

Economic Challenges. The economics of clinical trials present real challenges for medical device manufacturers and for society at large. By its very nature, clinical research is expensive to conduct. If new requirements for clinical data are imposed upon the device industry, there is some concern about whether the small companies that make up most of the industry will be able to afford the costs of conducting clinical research. Similarly, since budget pressures on the federal government show no signs of abating, there is concern about what will become of the funding for large-scale clinical work now being carried out by the National Institutes of Health, by other government agencies, and by university researchers supported through government grants.

It is not clear that society at large is willing and able to support ever more sophisticated and costly testing for medical interventions. The growing use of quality-of-life measurements as part of clinical test responses and the relatively recent inclusion of pharmacoeconomic end points in clinical trials are attempts to come to grips with this important question. Given these challenges, the need for efficiency in the conduct of clinical studies will become greater with time.

Although clinical trials are among the most useful methods for determining the safety and effectiveness of medical devices, there are dangers inherent in their use. If misunderstood or misused, the methods of clinical research can become a vehicle for callous disregard of the rights of the individual. Overreliance on clinical trials can also create unnecessary barriers to the advancement of medical science.

The dangers inherent in clinical research can be minimized. Manufacturers can ensure the protection of human subjects of clinical trials by observing the three principles of respect for persons, beneficence, and justice. Thoughtful and rational consideration of the concepts of clinical utility and relative efficacy can be used to determine a practical level of proof for showing safety and effectiveness. And statistical rigor in the design and expectations of clinical trials need not be a burden if it is applied in a way that accounts for the complexity of today's medical devices and their use.

Increasingly, new drugs and devices are being developed for world markets rather than for individual countries or regions. Groups such as the International Conference on Harmonization are developing consensus standards that will eventually be accepted by regulatory authorities in multiple countries. In time, such standards will lead to less duplication of clinical research efforts across international boundaries and, hence, more efficient clinical development of products for worldwide markets.

At the same time, increasing use is being made of modeling activities and small-scale pilot or feasibility studies. This practice promises to enable manufacturers to more rapidly identify medical interventions that have clinical efficacy and weed out less-promising therapies before large-scale clinical trials are performed.

As with any scientific tool, if used appropriately, clinical trials can aid the scientific process and the rapid introduction of safe and effective products. If not used correctly, however, their scientific requirements can become regulatory barriers.


Kshitij Mohan is a former director of FDA's Office of Device Evaluation; he is now corporate vice president for research and technical services at Baxter Healthcare Corp. (Round Lake, IL), and a member of the MD&DI editorial advisory board. Harold E. Sargent is director of applied statistics in the corporate research and technical services office of Baxter Healthcare Corp.

Who's Right in the FDA Numbers Game?

Medical Device & Diagnostic Industry Magazine | MDDI Article Index

Originally published January 1996

Regulatory Affairs

The House Commerce Committee has joined the chorus of those questioning the review productivity of FDA's Center for Devices and Radiological Health (CDRH), but its interpretation of the numbers may give the casual observer pause. On October 30, 1995, the panel's Subcommittee on Oversight and Investigations released a new study conducted by the congressional auditing agency, the General Accounting Office (GAO). The study concludes that FDA took a median of 804 days, or more than two years, to grant premarket approval (PMA) for Class III devices in 1993. This is almost double the agency's clearance time in 1989, it said. The 1989 median of 414 days was cause for great concern both in the agency and among industry members at the time, a House subcommittee staff document contended, since the Medical Device Amendments mandate FDA approval in 180 days after PMA application submission.

"The increased delays occurred even though the FDA received only half as many applications in 1994 as it did in 1989, and despite the fact that the agency's medical device program was increased by $20 million during fiscal year 1994," asserted Congressman Joe Barton (R­TX), chairman of the subcommittee. "More than any other development so far, this report...confirms what our hearings have found all year--that the FDA is delaying, not improving, Americans' access to medical treatment."

Barton charged that similar trends are occurring with 510(k) clearance times. According to the GAO report, 510(k) approvals took a median of 152 days in 1994, up almost 100% from the 80-day median in 1989. Similarly, the mean time for approval of these applications increased to 166 days in 1994, up from 124 days in 1989. The congressman added that the mean for 1994 will rise when the remaining 1994 backlog of applications is cleared.

In contrast, CDRH head Bruce Burlington has cited statistics that show "we were able to modestly decrease the average time it took FDA to review a PMA application from 21.5 months in fiscal year 1994 to 20 months in fiscal year 1995." He added that "we also cut by more than 50% the average time it took us to reach a final decision on a PMA after we reached an evaluation from one of our advisory panels."

CDRH has reduced the agency's average review time for 510(k)s by 24%, to 138 days in fiscal year 1995 from 182 days in fiscal year 1994, Burlington reported. Median total 510(k) time (from time of submission to final decision by the agency) similarly dropped in those two fiscal years, to 102 days in fiscal year 1995.

Obviously, the GAO is not counting days in the same fashion as FDA. So what accounts for the discrepancy, and who's right?

Two explanations emerge. First, the two studies reviewed different periods. Burlington used data from fiscal years 1994 and 1995, or a period from October 1, 1993, through September 30, 1995. The GAO, however, studied PMA applications in fiscal years 1989 through 1993, and 510(k)s in one additional year, 1994. Thus, for PMA applications, none of the periods covered in these studies overlap. For 510(k) reviews, there is only a little overlap.

However, CDRH data show that the median 510(k) review time steadily climbed in late 1992, surpassing 100 days in September of that year, and peaking at nearly 200 days in February 1993. The median time for this class of product reviews then began to drop and the decline continued throughout fiscal years 1994 and 1995.

From this comparison, it would seem that both studies could be correct--it's just that the GAO report is using older data. But the second manner in which the two studies differ is even more important in explaining the difference between them. It has to do with how review time is counted, or as Barton put it, "how the clock is stopped."

According to the GAO report, it counted applications and days "according to the fiscal year in which the applications were submitted to FDA. By contrast, FDA commonly reports review time according to the fiscal year in which the review was completed." Both studies counted total elapsed time (from the time a company submits its paperwork to the time FDA actually clears the product for market), including both the time the agency was waiting for more information from the submitting company and the time it spent reviewing the application.

Although Barton charged FDA with "statistical sleight of hand" for basing review-time calculations on the year of completion, the congressional report stated that both methodologies are perfectly valid ways of measuring FDA productivity. In fact, the GAO study even defended both methodologies: "using the date-of-decision [FDA's method] is useful when examining productivity and the management of resources," it said, while "using the date-of-submission [the GAO method] is useful when examining the impact of a change in FDA review policy." This discrepancy in methods is not, as the House subcommittee staff charged, a way of "manipulating the clock" in the agency's favor. In fact, it appears that FDA's response that the GAO report does not recognize the effect of review policy changes during the period studied is borne out by the differences in the two studies' data.

The House subcommittee's charges to the contrary, the apparent slide in CDRH review productivity is not due to some flaw in its structure or power (both of which the new Republican congressional leadership has proposed to scale back during the past year). Both studies generally agree that the climbing median review times occurred in the early 1990s--the time during which the agency was dealing with the extraordinary workload Congress handed it courtesy of the Safe Medical Devices Act of 1990. Moreover, the GAO study hints that the review times began to decline during the latter part of the period it studied, and does not report on fiscal years 1994 and 1995, when FDA's reported review times improved even further. In fact, Barton's charge that the $20 million had not improved productivity appears false in light of the FDA report's numbers, for these are the very years in which the extra money was actually spent.

The congressman is on firmer ground, however, when he discusses how the number of submissions from industry has fallen over the past several years. Indeed, both the FDA and GAO studies agree on that score, and if Congressman Barton is looking for a cause of this trend, perhaps he need look no further than the major changes to device law enacted in 1990.--Cliff Henke

Revolution by Phases: FDA's Regulation of Investigational Device Exemptions

Medical Device & Diagnostic Industry Magazine
MDDI Article Index

Jonathan S. Kahan

The collection and evaluation of clinical data to demonstrate the safety and efficacy of a medical device are absolutely essential for the ultimate premarket approval of all Class III devices. Under premarket approval (PMA) regulations, FDA requires the submission and review of valid scientific evidence to determine whether a reasonable assurance exists that the device is safe and effective, and has clinical utility.

As defined in the regulations, valid scientific evidence consists of data from well-controlled investigations, partially controlled studies, studies and objective trials without matched controls, well-documented case histories compiled by qualified experts, and reports of significant human experience with a marketed device.1 Notwithstanding the language of the regulations, FDA has interpreted the valid scientific evidence requirement as essentially limited to well-controlled investigations--which, as will be discussed, has led it to institute stringent requirements for the randomization and blinding of clinical trials.

Although clinical trial issues have historically been linked primarily to PMA submissions, they are also becoming more crucial to the marketing clearance of devices proceeding through the 510(k) premarket notification process. While it is difficult to obtain reliable information as to the number of 510(k) notices that are actually supported by well-controlled clinical data, FDA policies under Commissioner David Kessler have clearly led to requirements for greater data support for 510(k) notices, including increased requests for performance testing of Class II devices. Such performance testing can range from mechanical and other bench testing to animal and human clinical studies. It appears that the number of 510(k) premarket notifications supported by clinical trial data has jumped from 2­4% prior to 1990 to approximately 10% in 1995. This increase has led to a much more heightened awareness of clinical trial issues in the medical device industry.

The motivation for FDA's increased emphasis on clinical data can be traced directly to Commissioner Kessler's concern that the Office of Device Evaluation (ODE) in the Center for Devices and Radiological Health (CDRH) was not adequately evaluating the underlying clinical support for many device clearances. Kessler was especially concerned about the lack of clinical information and analysis relating to the ODE clearance of 510(k) notices for preamendment Class III silicone-gel breast implants. Consequently, he requested that Robert Temple, director of the Office of Drug Evaluation I at the Center for Drug Evaluation and Research (CDER), form a committee of CDER physicians and biostatisticians to evaluate the ODE clinical review process.

Under Temple's direction, the Committee for Clinical Review looked at numerous PMA applications, investigational device exemption (IDE) applications, and 510(k) premarket notifications to determine whether the underlying studies had been properly designed and conducted, and whether ODE had appropriately evaluated the trial data. In March 1993, the Temple Report was released.2 The committee found the following design deficiencies in the clinical data submissions received and evaluated by ODE:

  • Failure to specify a clear hypothesis to be tested and to develop a clear plan to test it.
  • Failure to enroll a sufficient number of patients to answer the primary study questions.
  • Failure to adequately specify requirements for patients entering studies.
  • Failure to identify a control group.
  • Failure to assess properly the comparability of patients in the treatment and control groups.
  • Failure to clearly and precisely define end points.
  • Failure to implement blinded evaluation of end points, especially when they were subjective in nature.

From March 1993 forward, the criticisms leveled by the Temple committee have been the driving force for most IDE clinical design decision making. As will be discussed later in this article, the entire IDE application process underwent a sea change in the period from 1990 through 1995. Indeed, some who have been involved in the process believe that ODE's thinking as to what is an appropriate design for the study of medical devices has undergone a revolution, not just evolution. With this background, it is possible to understand more fully the IDE process about to be described.

Clinical studies that are designed to evaluate the safety and effectiveness of a medical device are covered by the IDE regulations at 21 CFR 812. A sponsor of a clinical study must submit an IDE application to ODE for all studies that are intended to include a significant- risk (SR) device, such as an implantable pacemaker or an implantable orthopedic prosthesis. For non-significant-risk (NSR) devices, the sponsor is still required to follow the IDE regulations at 21 CFR 812, but no IDE application need be approved by FDA prior to the initiation of the clinical trial. Nevertheless, under the abbreviated IDE requirements for NSR devices, institutional review board (IRB) approval must still be obtained prior to initiation of the study, and requirements covering appropriate informed consent, recordkeeping, and adverse-effect reporting are still applicable.

Significant-Risk versus Non-Significant-Risk Device Studies. In initiating a clinical trial, one of the first steps that a study sponsor must take is to decide whether the device is SR or NSR. A significant- risk device is defined as one that:

1. Is intended as an implant and presents a potential for serious risk to the health, safety, or welfare of a research subject.

2. Is purported or represented to be for use in supporting or sustaining human life and presents a potential for serious risk to the health, safety, or welfare of a research subject.

3. Is for a use of substantial importance in diagnosing, curing, mitigating, or treating disease, or otherwise preventing impairment of human health, and presents a potential for serious risk to the health, safety, or welfare of a research subject.

4. Otherwise presents potential for serious risk to the health, safety, or welfare of a research subject.3

In order to assist sponsors in determining whether their device is SR or NSR, FDA issued a guidance document in June 1986 entitled "Guidance on Significant and Nonsignificant Risk Device Studies." This guidance document, which was updated in 1994, provides a list of examples of NSR and SR devices.4 In most instances, a sponsor will be able to make a fairly straightforward decision as to whether the device requires an IDE application because it is an SR device. However, even if a device is an NSR product, it is still best to review the study protocol with ODE even though no IDE application has been filed.

The rationale for discussing NSR device investigational plans with ODE prior to study initiation is that the office may have questions with respect to the design of the study, the success criteria and end points, the inclusion or exclusion criteria for the study, the sample size, or the case report forms. It is much better to have ODE input before beginning the study than to have FDA inform the sponsor at its conclusion that the study design was inappropriate and that, therefore, in order to obtain 510(k) clearance or premarket approval, the company must repeat the study under a new protocol.

Exempt Studies. In some cases, unapproved and uncleared devices are exempt from the IDE requirements. The most prominent of these exempt devices are the following:

  • Diagnostic Devices. A diagnostic device is exempt from the requirements of 21 CFR 812 if the testing is noninvasive, does not require an invasive sampling procedure that presents significant risk, does not by design or intent introduce energy into a subject, and is not used as a diagnostic procedure without confirmation of the diagnosis by another medically established product or procedure.5 In vitro diagnostics (IVD) that are exempt from the IDE requirements under 21 CFR 812 must still meet the requirements for appropriate labeling of "investigational or research use only diagnostics."6 A research-use-only device must be labeled "For Research Use Only. Not for use in diagnostic procedures."7 Similarly, investigational-use-only IVDs must be labeled "For Investigational Use Only. The performance characteristics of this product have not been established."
  • Consumer Preference Testing. A device is also exempt from the IDE requirements if it is undergoing consumer preference testing, testing of a modification, or testing of a combination of two or more devices in commercial distribution, if the testing is not for the purpose of determining safety or effectiveness and does not put subjects at risk. FDA takes a very narrow view as to whether a device is actually exempt from the IDE requirements because it is undergoing consumer preference testing. Indeed, if there is even a hint that the study is being conducted to determine the safety or effectiveness of the device, then the IDE requirements must be met.
  • Laboratory Testing. Devices shipped solely for research on or with laboratory animals and labeled "CAUTION--Device for investigational use in laboratory animals or other tests that do not involve human subjects" are exempt from the IDE requirements.
  • Custom Devices. Custom devices are also exempt from the IDE requirements. A custom device is defined as one that: (1) necessarily deviates from devices generally available or from an applicable performance standard or premarket approval requirement in order to comply with the order of an individual physician or dentist; (2) is not generally available to, or generally used by, other physicians or dentists; (3) is not generally available in finished form for purchase or for dispensing upon prescription; (4) is not offered for commercial distribution through labeling or advertising; and (5) is intended for use by an individual patient named in the order of a physician or dentist and is to be made in a specific form for that patient, or is intended to meet the special needs of the physician or dentist in the course of professional practice.8 Again, FDA takes a very narrow view as to whether shipment of an unapproved, uncleared device is in fact the distribution of a custom device not subject to the IDE requirements. The agency is presently developing a guidance document for industry relating to the appropriate definition of custom device. Suffice it to say at this time, however, that the shipment of any significant number of unapproved or uncleared devices under the custom device rationale may not be able to withstand FDA scrutiny.

Finally, there are several other exemption categories that merit only passing reference. Studies involving certain preamendment devices, devices found to be substantially equivalent, and veterinary devices are also exempt from the IDE requirements.

On-Line Resources Simplify Data Gathering

Medical Device & Diagnostic Industry Magazine | MDDI Article Index

Originally published January 1996

Information Technologies

Surfing the Internet or going on-line can be more than just a diversion; electronic information sources for the device industry report breaking news as well as pertinent regulatory affairs information. Instead of tracking down information through lengthy correspondence and faxes or wading through reams of report pages, as was once necessary, device industry personnel can now access current news almost instantaneously through on-line resources. Most services update daily and cull their information from a range of sources far wider than the average person alone could process, including FDA reports, government legislation, federal regulations, and corporate press releases. These services range in cost from expensive to free and generally require no more than a computer and modem to access.

Representative of the on-line proprietary services available are MDI Online and DeviceNet, two subscription services; a notable free service is the Regulatory Affairs Information Home Page, located on the Internet's World-Wide Web. Each provides useful industry, government, and regulatory affairs information. The proprietary services summarize information from many sources and update daily; the home page informs through links to other primary-source home pages. All can be accessed by IBM or Macintosh platforms.

The MDI Online package by Medical Data International, Inc. (MDI; Irvine, CA), offers access to industry news from several different sources. Three MDI publications, MedPro Month, Competitive Insight, and Inside Surgery, are available on the service. In addition, a daily on-line newsletter, Medical Industry Today, provides breaking industry news. "Medical Industry Today provides anywhere from 30 to 40 breaking stories each day geared toward the device, diagnostics, supply, and equipment industries," says Kevin Hicks, director of information technologies for MDI. "It's always up-to-the-minute, published each morning electronically."

In addition, the MDI Newsroom offers 22 different publications on-line, including Applied Genetics News, Health Legislation and Regulation, and Membrane & Separation Technology News. It also provides a federal news service with "same-day verbatim" transcripts of presidential statements; White House, state, and defense briefings; major congressional hearings; speeches, statements, and press conferences by leading policymakers and newsmakers; and appearances by major figures on television news shows, all transmitted within one or two hours of the event.

Subscribers can also access Internet newsgroups, arenas for discussion of various industry-related topics. "We have an easy-to-use front end, so someone who's not Internet savvy can easily use our system to send messages from the system into the Internet newsgroups," notes Hicks.

The basic service costs $4400 per year for unlimited use by one subscriber. However, the dial-up number is in the 714 area code, which will result in long-distance charges for nonlocal users. Subscribers will be able to log into MDI Online through local providers by the first quarter of next year.

The service requires a computer with a minimum of 4 MByte of RAM, 8 MByte of available storage on the hard drive, and a 9600-baud modem; Windows-based computers should have a cpu of 386 or better and a VGA monitor; Macintosh systems should be 7.0 or higher. For more information on MDI Online call Raymond Hecker, business development manager, at 714/251-2739 or send E-mail to ray@medicaldata.com.

DeviceNet, by Schiff & Co. (West Caldwell, NJ), also provides device industry and regulatory news, updated daily. The package divides the information into industry news, Federal Register, 510(k) listings, warning letters, and FDA enforcement reports. All of the data can be downloaded or searched.

All information presented by DeviceNet is filtered daily from various news wires, FDA sources, and DeviceNet staff research. "The regulatory information comes directly from FDA. Nine times out of 10 we will have it on our system before it's available in print," says DeviceNet's designer and programmer Jack Parker, vice president of the food and cosmetic division of Schiff.

"We will not just take a press release off the wire and send it through. There is a screening process. We encapsulate the news, so if there is a press release that's two or three pages long, you generally will get no more than a headline and a paragraph of the distilled information."

The system was originally designed for use by regulatory affairs staff in the industry. "I quickly found out that, although lots of regulatory affairs people are on the service, we get at least as many R&D, sales and marketing, and executive users," says Parker.

Potential subscribers may test DeviceNet free for 30 days, dialing up with either Schiff's software, SchiffCom (IBM format only), or with any communications software (Mac or PC).

The system can function with any modem speed above 300 baud and can be accessed by any communications software. Schiff will offer a World-Wide Web version soon. For more information on DeviceNet, call 201/ 227-1830 or send E-mail to schiffco1@ aol.com.

DeviceNet costs $595 per year plus possible long-distance phone charges on the access number. Parker says that not having a toll-free access number keeps the cost of the service down. "We looked at what competitive print information and electronic sources cost, and we priced DeviceNet at generally less than half."

Free regulatory affairs information is available on the Regulatory Affairs Information Home Page on the Internet's World-Wide Web. Don Kafader put the page up last February as an extension of his work as manager of regulatory affairs at Organon Teknika Corp. (Durham, NC) and as a consultant. The page contains links to more than 400 different sites and documents that relate to regulatory affairs, including FDA sites. Kafader maintains and updates the site himself. "I have a personal interest in making information available. The agencies and the guidelines exist, and this is a perfect vehicle to extend that to, basically, the world," he says.

Kafader gathers information for his site from feedback of visitors to it and from monitoring Internet newsgroups. He also checks the various web search engines, such as Yahoo and Web Crawler, for new sites. "Whenever I become aware of a site there's nothing really that prevents me from having a link there 30 minutes later if I am so inclined," says Kafader. "Hopefully I am very current."

The linked sites include information from the Federal Register, U.S. legislation, the Code of Federal Regulations, professional organizations, and consulting organizations. There were about 600 hits, or visits, to the home page itself during the week of November 12, 1995. In the same week, however, there were over 1000 hits to the Federal Register page, which can be accessed through the home page. "That is the page that has been getting the most activity consistently for the last three or four months," adds Kafader.

The home page address is http://www. nando.net/ads/ckbus/RAinfo/reglink1.htm. Kafader also maintains a bulletin board system (BBS), the Regulatory Forum BBS, which can be accessed at 919/848-9461.

Gathering information through on-line services is efficient in terms of time and energy saved; it is also a paperless process. These services are the equivalent of the daily newspapers, with the added advantages of continuous updates, instantaneous transmission, and the ability to search out specific information. They act as a personal research assistant, in some cases at considerably less cost than that of hiring a staff researcher. As activity on the World-Wide Web increases, more resources will appear, making it easier to stay up-to-date on industry happenings--and, perhaps, challenging proprietary services to stay one step ahead of the free ones.--Sashi Sabaratnam

Continuation of MDDI's January 1996 article, "Revolution by Phases: FDA's regulation of Investigational Device Exemptions"

Back to previous page

Although it is probably belaboring the obvious, the design and conduct of the clinical trial are the keys to ultimate marketing approval of the device. The data generated by the clinical trial must support the claims that the sponsor intends to make for the marketed device and the clinical utility of the device. In other words, if the sponsor wishes to market the device for use in a specific study population, that population must be included in the clinical trial. Similarly, if the ultimate instructions for use of the product will recommend a specific power setting or duration of use, these must be evaluated in the clinical trial. If prior to initiating the clinical trial the company is unsure as to what settings or duration of use are appropriate, then a feasibility study may be called for. ODE has been encouraging companies to conduct "pilot" or "feasibility" studies prior to the initiation of clinical trials to resolve such issues as device design and operating specifications and to better define patient populations and device indications.

ODE has issued a guidance document regarding feasibility studies which is entitled "Guidance on the Review of Investigational Device Exemptions (IDE) Applications for Feasibility Studies."12 Under that guidance, an IDE application for a feasibility study must address all of the elements of a standard IDE application. However, the study is typically conducted with a smaller number of patients, often 10 or fewer and rarely more than 20. Feasibility studies are often conducted at only one or two institutions with very specific end points in mind. For example, if a software algorithm is necessary for the appropriate operation of the device, the company may wish to evaluate and finalize that algorithm in a limited study rather than proceed to a trial with an algorithm that is uncertain with respect to ultimate safety or efficacy. Once a feasibility study is completed, the company can often then submit an IDE supplement to expand the study to the pivotal clinical trial.

After the company has made the decision to proceed to the pivotal clinical trial, trial design becomes of foremost importance. In September 1993, FDA issued a guidance document entitled "Medical Device Clinical Study Guidance," in which it covers in great detail the issues that should be considered in the design of a clinical trial.13 It should be mentioned that many of the recommendations in this guidance document can be traced directly to concerns enunciated in the Temple Report.

Without going into the clinical study guidance document in great detail, it contains several points worth mentioning. First, it appropriately states that good clinical study design controls minimize known or suspected sources of bias and other errors so that clinical device performance can be assessed clearly. One of the primary concerns of ODE since release of the Temple Report has been the need to reduce potential study biases that result when any characteristic of the investigator, study population, or study conduct interferes in any way with the ability to measure a variable accurately.

Second, the guidance document emphasizes that the study population and the control population must be well-defined. It discusses active concurrent controls, passive concurrent controls, self controls, and historical controls. It mentions, however, that active concurrent controls and, where applicable, self controls allow the greatest degree of opportunity for comparability.

Experience over the past several years has revealed that active concurrent controls with appropriate randomization and blinding are the key elements FDA considers for appropriate clinical design. While almost all PMA applications prior to 1990 were approved based upon comparison of the study results to historical controls, FDA now accepts historically controlled studies only in the rarest of instances. Partly because of the influx of CDER personnel into CDRH, the drug model for clinical trials--which is grounded upon randomized, double-blinded, and controlled trials--has become the gold standard for medical device studies as well. The September 1993 guidance document makes it very clear that "historical controls are the most difficult to assure comparability with the study population" and "will usually require much more work to validate comparability than concurrent controls." Therefore, unless absolutely necessary for ethical or other reasons, historical controls should be relied upon sparingly in proving the safety and efficacy of a new device.

Studies for IVDs also raise quite different issues from studies for other medical devices. First, very few IVD clinical trials are conducted under the IDE regulations. Because they involve exempt diagnostics, they typically proceed without the filing of an IDE application. Nevertheless, because of the importance of the clinical study data to the ultimate clearance or approval of the device, most IVD manufacturers review their study protocols with the ODE Division of Clinical Laboratory Devices (DCLD) prior to initiation of their clinical trials.

DCLD has many guidances that apply to specific diagnostics, and they should be reviewed very carefully by IVD study sponsors. For example, the guidance document "Review Criteria for Premarket Approval of In-Vitro Diagnostic Devices for Detection of Antibodies to Parvovirus B19" contains a very detailed section on clinical studies.14 It discusses how clinical confirmation of infection may be conducted, how the test sample should be submitted to the clinical laboratory, and the bases for clinical diagnosis.

DCLD has also issued a general guidance document on the conduct of IVD clinical trials which should be reviewed by sponsors prior to conducting such trials. Entitled "Points to Consider for Collection of Data in Support of In-Vitro Device Submissions for 510(k) Clearance," it covers study protocols, sampling methods, study site requirements, product inserts, and the responsibilities of principal investigators of clinical trials.15 This guidance document has not been finalized and contains several controversial provisions, including the requirement that the investigator sign off on the study, indicating that he or she has reviewed and verified the data and the manufacturrer's presentation of the data analysis to FDA. Nevertheless, this guidance is helpful in evaluating how DCLD looks at the important clinical issues for IVD trials.

Finally, for IVDs and all other devices, the issue of demonstrating clinical utility as well as safety and efficacy should not be overlooked. FDA has recently been placing a greater emphasis on demonstrating clinical utility in device trials, much to the consternation of the device industry.

Device Company Finds Savings in Installation Procedures

Medical Device & Diagnostic Industry Magazine | MDDI Article Index

Originally published January 1996

Management Strategies

For device manufacturers, the key challenge of the decade may well be to satisfy the increasing needs of their customers to shave costs while improving service. For most companies, this challenge means looking beyond traditional improvements in design and manufacturing to areas such as distribution and service. A recent example is Varian Oncology Systems (Palo Alto, CA), which has found fertile ground for innovation in its installation process.

Six years ago, Varian took almost 10 weeks (500 hours) to install one of its 9-ft-tall, 10- to 15-tn medical linear accelerators used in the treatment of cancer. Installation costs ran in the millions of dollars annually. Since then, the company has cut installation time in half to five weeks (215 hours), reduced installation costs by 41%, and simplified the installation process by 30%.

Varian accomplished this by no high-tech means. Instead, it employed cross-functional teams of service and manufacturing personnel to brainstorm and analyze the installation process, aided by process-mapping software. These teams uncovered and corrected wasteful steps during installation, and rectified a pattern of inadequate communication between the field and manufacturing personnel.

"Varian Oncology Systems first used these continuous-improvement approaches in its manufacturing, and has applied them to improve all its business processes," says Keith Krugman, vice president for customer support. Varian set as a goal the reduction of excessive time and costs in installing its accelerators. To tackle the problem, the company's customer-support organization established the cross-functional teams to examine the installation process and its many loopholes. The teams charted the installation process, from start to end, using a process-mapping software application, Visio (Sharpware Corp.), to create detailed flowcharts of the process.

Initial discussions uncovered some 100 issues, most of which were unnecessary steps in the installation process. Though each solution saved only an hour or two of installation time, combined they saved many more. Early savings totaled $750,000.

Other, more product-related issues necessitated altering the design of Varian's linear accelerator or its packaging. For instance, the unit's klystron, or x-ray energy source, was previously shipped separately to sites. Today, additional bracing devices hold the klystron in place, so it can be shipped as part of the accelerator. This change saved some five hours per installation.

In addition, the accelerator's modulator doors and emergency off-switches are now installed at the factory, saving an estimated two hours per installation. Similarly, pilot holes were added to the unit to help reduce by one hour per installation the time needed to pin the equipment together.

The Varian teams saved additional installation time by changing packaging methods and shipping with the units materials needed for installation. They determined, for instance, that the foam packaging "popcorn" used to protect some accelerator parts was unnecessary, merely creating a mess that service people had to sweep up after the machine was installed. Eliminating the popcorn saved at least one hour per installation.

Further, they found that hardware was often misplaced on-site, and that service staff had to go to a hardware store for the necessary nuts and bolts. Now, the accelerator shipment includes a spare parts kit. This change saved some two hours per installation. Similarly, they found that service staff had to scout for a grocery store to purchase the distilled water needed by the accelerator. Today, distilled water is packaged with the accelerator, saving one to three hours per installation.

In addition to improving the productivity of its installation teams (the same number of installers can do 75% more installations than in the past), Varian has improved its service to customers. According to Varian's Krugman, "The improvement has helped several hospitals open their radiation therapy departments on schedule and gain a faster return on investment."

As health-care providers become increasingly cost-conscious, more device companies may follow Varian's lead in wringing inefficiencies from the installation process. --Robert Seeley